Interpretable Assessment of ST-Segment Deviation in ECG Time Series

Figures

• 

  Figure 1. ECG beat morphology example. Note the vertical dotted lines representing the ST-segment, which is the interval between ventricular depolarization (shift in electric charge distribution inside the cell) and repolarization (J-point elevation) [4].

• 

  Figure 2. ST-segment deviation visual, the traditional way to identify this abnormality by measuring the distance between the ST-segment and the isoelectric line. Normal ECG beat (left), elevated ST-segment (center), depressed ST-segment (right).

• 

  Figure 3. Overview of the process to the identification of ST-segment deviations.

• 

  Figure 4. Confusion matrix on the modeling containing the features concerning the distances between the waves of the ECG beat.

• 

  Figure 5. Confusion matrix on the modeling containing the features concerning statistical measurements, SQIs of the ECG beat and distance from the baseline at two different points after the J-point, 60 and 80 mV respectively.

• 

  Figure 6. Confusion matrix on the modeling containing the features concerning statistical measures, distance from the baseline to J-point, and SQI of the beat ECG.

• 

  Figure 7. Confusion matrix on the modeling containing the features concerning statistical measures and SQIs of the ECG beat.

• 

  Figure 8. The area under the curve with binarization of the problem, considering only normal and abnormal beats (depression or elevation as one).

• 

  Figure 9. MLP results for ST deviation calculation in ECG time series.

• 

  Figure 10. ELM results for ST deviation calculation in ECG time series.

• 

  Figure 11. LSTM results for ST deviation calculation in ECG time series.

• 

  Figure 12. CNN results for ST deviation calculation in ECG time series.